Making the right choices in a disaster

INTAMAP map showing measurements of Nitrogen dioxide across the UK

Example INTAMAP interpolation (centre) of Nitrogen dioxide measurements across the UK

If there were a nuclear accident in Europe, how would those responsible for disaster response know where to focus attention? And how would they access the right information to make fast and accurate decisions?

Aston University are collaborating with eight other European partners to deliver the Interoperability and Automated Mapping Project (INTAMAP), which will provide emergency response decision makers with the complex data and information they need over the internet in a usable format… fast.

So, in the case of a nuclear accident, for example, they could use INTAMAP to make the right decisions quickly.

The existence of the internet has revolutionised our ability to gather data from around the globe and make decisions based on a wide variety of observations. Through the internet, huge and diverse datasets can be pulled together, whether for comparing insurance quotes or assessing tornado damage in a remote area.

Decision makers don’t just need raw numbers though – they need to be able to analyse and combine different pieces of (potentially conflicting) evidence in order to make the best choices.

For example, data from radiation monitoring sensors across the whole of Europe might be quickly gathered in case of a nuclear accident, in order to work out which areas exceed the danger threshold, and should thus be evacuated.

However, sensors are expensive to install and maintain, so there will be many locations at which an accurate radiation value is not known. In this case, it’s necessary to use powerful algorithms to interpolate between the data points. In other words, to try and predict the radiation levels and associated risks at locations that aren’t being monitored.

Traditionally, such an interpolation would be done using a specialised software application installed on a desktop but recent developments in web technology mean that such applications can be accessed remotely, through standardised mechanisms known as “web services”.

The application itself is located on a web server (just like an email application or an online database), and can be called from any other computer on the network (using HTTP, the standard internet protocol).

So, for example, if Photoshop was exposed as a web service, and you have the URL (internet address) for that web service, you can access and use complex image editing functions over the web without the need to install or upgrade any desktop application.

In the radiation example, the locations and radiation levels from the sensor network could be sent to an interpolation web service from any computer, and a map of likely radiation values covering the whole of Europe could be sent back.

The INTAMAP project is about allowing just this sort of fast and reliable interpolation, by using web services to expose complex algorithms based on Bayesian statistical approaches, which might not usually be accessible to decision makers. The algorithms are made available to any device that is connected to the internet. The sensor readings themselves might also be supplied by a web service, so it’s possible to use the outputs of a chain of web services without having to have a powerful machine or expensive licensed software in front of you.

One thing you may have noticed about the example of a radiation emergency is that the system produces maps of ‘likely radiation’. This might not sound particularly encouraging in an emergency situation, since surely we want an accurate picture? But, paradoxically, recognising that information will never be perfect allows safer decisions to be made, since the effects of chance and uncertainty can be properly addressed.

All data is uncertain, whether because of faults and biases in the measuring equipment, or because—in the case of interpolation—a value has had to be inferred from nearby points. If a user has a good idea of how reliable the data is, (say, how well a sensor usually performs) they can decide on the appropriate level of caution, generating best and worst-case scenarios in order to decide policy and action. Value can be added to the predicted maps by recording how reliable the estimate of radiation is at every location. As part of the INTAMAP project, a language called UncertML has been developed, which allows detailed information on data uncertainty to be exchanged and communicated automatically between machines and Web Services.

The INTAMAP project is now in its third year, and has brought together significant expertise from across Europe. There is growing interest in industry both in web services and in how a good idea of uncertainty can lead to more sensible and robust decisions.

Aston University plan, through further research and collaboration with a range of users, to continue building on the two real innovations of the project:

  • Opening up of high quality analysis functions over the web
  • Clear communication of information and its uncertainty for a variety of real-world applications, such as weather mapping, agricultural planning and contaminated land survey.

Thanks to Lucy Bastin and Dan Cornford from Aston University for providing us with the content for this post. For more information, contact Matthew Williams.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: